Why The Back Surface Matters After Dmek+Ce/Iol: Outcomes And A Practical Aiming Rule

To evaluate refractive outcomes after DMEK+CE/IOL, identify biometric predictors, and quantify corneal changes (anterior vs posterior keratometry, thickness metrics). We assessed correlations between postoperative MRSE and relative pachymetry and the B/F ratio (pre/post), Δ‑based correlations versus the per‑eye IOL target, and we introduce a cohort‑derived aim‑refinement formula.

Single center; N = 51 eyes; mean age 73.6 ± 5.8 years (median 75); median follow‑up ≈ 186 days.

Retrospective analysis of pre/post subjective refraction (MRSE), prediction error (Hoffer/Haigis/SRK/T), corneal parameters (Kmfront/Kmback, cylinders, pachymetry, relative pachymetry, B/F ratio). Pre/post comparisons per t‑tests in the dataset; Pearson correlations for MRSE vs relative pachymetry and B/F (pre/post), and Δ‑based correlations vs per‑eye IOL target. A practical linear aid maps expected ΔKmfront/ΔKmback and WTW to a predicted deviation from the base aim (PE).

Refractive outcome & prediction error: Postoperative MRSE ≈ -0.05 D (NS vs pre‑op; p = 0.312). Myopic prediction bias persisted across formulas (Hoffer -0.55 D; Haigis -0.65 D; SRK/T -0.55 D).

Corneal changes (pre vs post): Kmfront 43.66 → 43.33 D (NS; p = 0.411); Kmback -5.93 → -6.40 D (significant steepening; p ≈ 2.28e-06); pachymetry 600.8 → 522.3 µm; relative pachymetry 8.02% → -3.98%.

Performance of the aiming framework (quantitative claim) In‑cohort evaluation (n = 51): relative to the unadjusted plan (base aim), the intercept‑calibrated linear aid reduced RMSE from 0.944 D to 0.855 D (−9.4%) and MAE from 0.689 D to 0.643 D (−6.6%).

DMEK+CE/IOL produced MRSE close to target but with a persistent ~0.6 D myopic bias across standard formulas. Posterior corneal steepening after DMEK, with minimal anterior change, appears to drive part of the error. A cohort‑derived aiming framework can reduce overall error (≈9% RMSE reduction with intercept calibration), but posterior‑aware keratometry (EKR/TK/total keratometry) and constant optimization remain advisable. Importantly, the magnitude of postoperative posterior steepening (ΔKmback) exhibited strong baseline associations with preoperative B/F ratio (r = -0.82) and relative pachymetry (r = -0.69), indicating that more edematous corneas and higher B/F tend to undergo greater posterior corneal change after DMEK.